BCFD Predictions for the AIAA Drag Prediction …...Title Results_PW.ppt Author Joseph Morrison Created Date 6/27/2006 3:27:20 PM

BT_PW_no-icon_simple.ppt | 8/4/2006BOEING is a trademark of Boeing Management Company.Copyright © 2006 Boeing. All rights reserved.

BCFD Predictions for the3rd AIAA Drag Prediction Workshop(DPW3)Chad Winkler, Andy Dorgan, Mark Fisher, Mori ManiThe Boeing CompanySt. Louis, MO

S. P. VankaUniversity of Illinois at Urbana-ChampaignUrbana, IL

Boeing Technology | Phantom Works

Copyright © 2006 Boeing. All rights reserved.

BCFD Code Details

•Cell-centered, finite-volume approach

•HLLE flux calculation with second-order spatial reconstruction

•Linear preserving gradient calculation

•Fully implicit time integration

•Turbulence models

•Spallart-Allmaras

•SST

•Additional capabilities: Time accurate LES, real gas effects,hybrid structured/unstructured solver, additional fluxformulations available



Grid Details

• Unstructured grids

•Mixed tetrahedra and prisms (boundary layer)

•Surface grids generated with MADCAP

•Volume grids generated with AFLR3

•Available on NASA FTP site

•Running on 64 bit Linux clusters

•Typical execution time : 24 hours on fine grid (33M cells)running on 33 processors



F6 Wing/Body Grids

Coarse (~4M cells) Medium (~8M cells)

Fine (~33M cells)



F6 Wing Root Region Grid

Coarse (~4M cells)

Medium (~8M cells)

Fine (~33M cells)



F6 + FX2B Wing Root Region Grid

Coarse (~4M cells)

Medium (~8M cells)

Fine (~33M cells)



Wing Root RegionSurface flow – Fine grid CL=0.5

•No separation seen on theF6+FX2B geometry wing root

•Separation seen on the F6geometry wing root



CD - CL^2/(PI*AR)

CL

0.017 0.018 0.019 0.02 0.021 0.022 0.023 0.024 0.0250.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

F6+FX2B S-A

F6+FX2B SST

F6 S-A

F6 SST

Drag Polars

•Error bars represent magnitude ofoscillations of CL in the F6 solution

•F6+FX2B solutions saw littleoscillation

•SST model seen to predict ~10counts less drag than the S-A model

•FX2B fairing seen to reduce dragregardless of turbulence model



Skin friction behavior

CD_Skin Friction

CL

0.01 0.011 0.012 0.013 0.014 0.0150.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

F6+FX2B S-A

F6+FX2B SST

F6 S-A

F6 SST

•Lower SST drag comes fromreduced viscous dragcontribution



Grid convergence study

GRIDFAC = 1/(GRIDSIZE)^2/3

CD-CL^2/(PI*AR)

0 1E-05 2E-05 3E-05 4E-05 5E-050.017

0.018

0.019

0.02

0.021

0.022

0.023

0.024

F6+FX2B S-A

F6+FX2B SST

F6 S-A

•SST results seen to extrapolate toa lower drag value when comparedto S-A for the FX2B configuration



Crinkle cut, F6+FX2B , S-A , Mach contours atBL=200mm

Isotropic tetrahedraquickly dissipate wake



Wing Cp contours, F6+FX2B, S-A model

AoA = -3 -1 0 1.5

Top View

Bottom View



Comparison of Cp between turbulence modelsCL=0.5, fine grid

(Cp_Spalart – Cp_SST)

Cp cut at BL=240.37mm

x/c

Cp

0 0.2 0.4 0.6 0.8 1

-1

-0.5

0

0.5

F6+FX2B, S-A

F6+FX2B, SST

Shock moves forwardin S-A solution

αS-A=0.119°αSST=0.166°



Comparison of skin friction between turbulence models

(Cf_Spalart / Cf_SST)

CL=0.5, fine grid

•Localized regions of higher skin friction using S-A when compared to SST

αS-A=0.119°αSST=0.166°



Summary

•Strong need for best-practices in unstructured grid generation – bothsurface and volume gridding

•Refine wake region using localized source nodes in volume gridgeneration

•Difficulty converging F6 cases (without fairing) for both turbulence models

•Turbulence model + grid dependencies•~10 counts drag difference predicted between S-A and SST models

•Refine grid further to remove any grid dependency on turbulence model

•Future plans•Alternate grids – highly resolved and selectively resolved grids, other DPW3grids

•Unsteady simulations

•Cross-code solution comparisons

BT_PW_no-icon_simple.ppt | 16Copyright © 2006 Boeing. All rights reserved.

Documents

BCFD Predictions for the AIAA Drag Prediction …...Title Results_PW.ppt Author Joseph Morrison Created Date 6/27/2006 3:27:20 PM